AI captions and translations

Every session has live captions in the speaker's language and text translation into every language your audience picks. You do not switch them on.

What your audience gets

• Captions of what is said, in the session's Speaker language.
• Text translation in every language. Each listener picks their own language when they arrive.
• If the speaker's language has no speech recognition, the session has no captions. The session's page says so under Ready to go live?, and listeners can still get interpretation.

Help the engine get your words right

The Captions & translation section of the session form holds the settings that help:

• Event type tells the engine what kind of event it is, so captions and translations use the right words.
• Vocabulary lists names and terms the engine should listen for or correct.

See Glossary and context. Choosing the accent closest to the speaker's also helps: see Languages.

Line captions up with a livestream

If people at home watch a livestream next to the captions, set Caption delay for viewers at home under Video & replays. Match your livestream's delay, so captions don't appear before the speaker on video. It can be between 0 and 360 seconds.

Songs

Turn on Detect songs under More options to tag the captions sung during a song, and tell viewers once the room starts singing.

Check it worked

On the session's page, Ready to go live? shows Captions from the speaker's language into every language.